https://bugs.winehq.org/show_bug.cgi?id=50169
Bug ID: 50169 Summary: Regression: Microsoft Flight Simulator X hangs on startup (bisected) Product: Wine Version: 5.22 Hardware: x86-64 OS: Linux Status: UNCONFIRMED Severity: normal Priority: P2 Component: -unknown Assignee: wine-bugs@winehq.org Reporter: czbd@o2.pl CC: julliard@winehq.org Regression SHA1: 871d9863bc2d1c9406b47aa644974768d5c6d2c4 Distribution: ArchLinux
Created attachment 68686 --> https://bugs.winehq.org/attachment.cgi?id=68686 Git bisect result
With the 5.22 release of Wine, Microsoft Flight Simulator X has ceased to launch correctly - instead, it hangs on startup.
I attach the git bisect result (of regular Wine), a log taken with Wine Staging 5.21 in which I launch FSX and then close it as soon as it properly loads into main menu, and a log taken with Wine Staging 5.22.
https://bugs.winehq.org/show_bug.cgi?id=50169
--- Comment #1 from Czcibor Bohusz-Dobosz czbd@o2.pl --- Created attachment 68687 --> https://bugs.winehq.org/attachment.cgi?id=68687 Wine Staging 5.21 log
https://bugs.winehq.org/show_bug.cgi?id=50169
--- Comment #2 from Czcibor Bohusz-Dobosz czbd@o2.pl --- Created attachment 68688 --> https://bugs.winehq.org/attachment.cgi?id=68688 Wine Staging 5.22 log
https://bugs.winehq.org/show_bug.cgi?id=50169
--- Comment #3 from Czcibor Bohusz-Dobosz czbd@o2.pl --- Weird - I have just now built the preceding commit, just to make sure, and it also suffers from the same issue. I've also noticed that, for some reason, several sources in the wine directory have been untracked (they definitely weren't this way before the bisecting)...
So, apologies - looks like my git somehow got borked? Gotta bisect this one again...
https://bugs.winehq.org/show_bug.cgi?id=50169
Alistair Leslie-Hughes leslie_alistair@hotmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |leslie_alistair@hotmail.com Ever confirmed|0 |1 Status|UNCONFIRMED |NEW
--- Comment #4 from Alistair Leslie-Hughes leslie_alistair@hotmail.com --- (In reply to Czcibor Bohusz-Dobosz from comment #3)
Weird - I have just now built the preceding commit, just to make sure, and it also suffers from the same issue. I've also noticed that, for some reason, several sources in the wine directory have been untracked (they definitely weren't this way before the bisecting)...
So, apologies - looks like my git somehow got borked? Gotta bisect this one again...
The regression is caused by the disabled patchsets in staging, namely widl-winrt-support windows.gaming.input-dll windows.globalization-dll windows.media.speech.dll windows.networking.connectivity.dll
Some of the patchset in widl-winrt-support, are currently being upstream.
https://bugs.winehq.org/show_bug.cgi?id=50169
--- Comment #5 from Czcibor Bohusz-Dobosz czbd@o2.pl --- (In reply to Alistair Leslie-Hughes from comment #4)
(In reply to Czcibor Bohusz-Dobosz from comment #3)
Weird - I have just now built the preceding commit, just to make sure, and it also suffers from the same issue. I've also noticed that, for some reason, several sources in the wine directory have been untracked (they definitely weren't this way before the bisecting)...
So, apologies - looks like my git somehow got borked? Gotta bisect this one again...
The regression is caused by the disabled patchsets in staging, namely widl-winrt-support windows.gaming.input-dll windows.globalization-dll windows.media.speech.dll windows.networking.connectivity.dll
Some of the patchset in widl-winrt-support, are currently being upstream.
That would be even weirder, if I'm understanding right, because sure enough, after a second failed bisect (and a suspicion that it's actually caused by `CFLAGS="-g -O0"`; now running a third, it'll take a while), I have also checked regular, non-staging Wine 5.21 and 5.22 as packaged by Arch Linux - Wine 5.21 starts FSX just fine, 5.22 hangs.
Guess that my next step is obtaining less contaminated logs and sending 'em here ASAP. :)
Thank You!
https://bugs.winehq.org/show_bug.cgi?id=50169
--- Comment #6 from Czcibor Bohusz-Dobosz czbd@o2.pl --- Created attachment 68696 --> https://bugs.winehq.org/attachment.cgi?id=68696 Wine 5.21 log
Here we go. (The DirectX dll creation errors are because I have DXVK, however specifically for fsx.exe, all dlls are explicitly set to builtin.)
https://bugs.winehq.org/show_bug.cgi?id=50169
--- Comment #7 from Czcibor Bohusz-Dobosz czbd@o2.pl --- Created attachment 68697 --> https://bugs.winehq.org/attachment.cgi?id=68697 Wine 5.22 log
And it seems that raw Wine 5.22 actually displays this nice Don't Send Window™, something that I had omitted before, perhaps as a result of the combined "doesn't work? check" routine, and the irrational thought that maybe once in a blue moon (rather than consistently in a particular version) it just crashes instead of hanging... Definitely a humbling experience right here in this bug.
https://bugs.winehq.org/show_bug.cgi?id=50169
--- Comment #8 from Czcibor Bohusz-Dobosz czbd@o2.pl --- Another factor that contributed to my confusion is that what I am trying to bisect consistently hangs - and the latest finding is, with release 5.21 too. Time to consult the Arch PKGBUILDs and build it yet again, maybe these special flags they use are really mandatory on these new compiler versions over the ol' good ./configure --disable-tests --enable-win64 and a corresponding one for wine32...
Until this insanity is fixed, only these /usr/bin/wine logs are reliable out of what I submitted, I guess.
https://bugs.winehq.org/show_bug.cgi?id=50169
--- Comment #9 from Czcibor Bohusz-Dobosz czbd@o2.pl --- TL;DR By 9faa5eeddd24a057d9ff522259c9dbdc6203c098, it's apparently fixed, thanks!
After finding that CFLAGS="-march=x86-64 -mtune=generic -s -O2 -pipe -U_FORTIFY_SOURCE -D_FORTIFY_SOURCE=0" CROSSCFLAGS="-s -O2" CPPFLAGS="-D_FORTIFY_SOURCE=2" LDFLAGS="-Wl,-O1,--sort-common,--as-needed,-z,relro" consistently make Wine 5.21 builds that correctly run FSX, but still failing to bisect (the new attempt pointed to be8a81b54c7adf1e77d9411d337b9227e0eb119f, which, in two following separate manual builds from a clean state, came out working in one, broken in the other... I don't know what to think of this), I have just tried building the latest commit at this time (9faa5eeddd24a057d9ff522259c9dbdc6203c098), and FSX consistently loads with several rebuilds of that one.
So, yeah, the bug is fixed, it seems. And still is the weirdest software malfunction I've seen in a long time. As well as a good lesson for me not to put a status like "(bisected)" in the title of a bug report again... :)
Great thanks to everybody working on Wine, a difficult and truly important job.
https://bugs.winehq.org/show_bug.cgi?id=50169
Robert Wilhelm sloper42@yahoo.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |sloper42@yahoo.com Status|NEW |RESOLVED Resolution|--- |FIXED
--- Comment #10 from Robert Wilhelm sloper42@yahoo.com ---
(In reply to Alistair Leslie-Hughes from comment #4)
The regression is caused by the disabled patchsets in staging, namely widl-winrt-support windows.gaming.input-dll windows.globalization-dll windows.media.speech.dll windows.networking.connectivity.dll
Some of the patchset in widl-winrt-support, are currently being upstream.
The patchset was enabled again in
https://github.com/wine-staging/wine-staging/commit/3553024da0b61b3188c11696...
As Czcibor pointed out in comment 9, it's apparently fixed now. I will set bug status to resolved.
Alistair, thank you for all your wine-staging work. Czcibor, thank you for your bisecting work.
https://bugs.winehq.org/show_bug.cgi?id=50169
Alexandre Julliard julliard@winehq.org changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |CLOSED
--- Comment #11 from Alexandre Julliard julliard@winehq.org --- Closing bugs fixed in 6.0-rc1.
-
WineHQ Bugzilla